Following tasks are remaining from 2a review. Either complete for 2b or bump to later phase if appropriate

config

• Check why CCMMF_DIR is defined in .Renviron, other vars are defined in 000-config.R; update these files and explain in docs/workflow_documentation.md under 'Configuration Setup'
  • https://github.com/ccmmf/downscaling/pull/1/files#r2165513286

downscaling_results.qmd

• Check that county stocks are calculated correctly
  • What's driving the super-high AGB in San Joaquin Co? https://github.com/ccmmf/downscaling/pull/1/files#r2164957054
  • Why is Fresno county's total stock so much higher than Kern's when their density is similar and Kern is bigger? https://github.com/ccmmf/downscaling/pull/1/files#r2164953773
• See if environmental-characteristics-of-each-cluster plot can be improved

010_prepare_covariates.R

• Replace /projectnb/dietzelab/dongchen/anchorSites/NA_runs/soil_nc with files in data_dir, 'soilgrids/soilgrids_california_*'
• Consider using soilgrids 0-30 cm instead of 0-5 cm; requires aggregation

clustering

• write out cluster environmental chars + file mapping site_id <--> cluster
• identify 'most representative sites' from each cluster (for targeted multisite sensitivity and other analyses)

Phase 3

• Clean up 009_update_landiq.R
• query rasters w/ polygons in 010_prepare_covariates.R
• Update design points (existing issue I think)
  • at that point, update 030_extract_sipnet_output.R to read site info from settings file.
  • that will also be a good time to consider contributing the function that goes from pecan outdir --> ensemble_output.csv to PEcAn, along with the functions that take ensemble_output--> array and --> nc
• remove "DIRECTORY AND FILE CHECKS section from 030_extract_sipnet_output.R; salvage anything useful for tests.
• Consider using optparse